Title The mitogenome of a Malagasy Malaza fastuosus (Mabille, 1884) recovered from the holotype collected over 140 years ago adds support for a new subfamily of Hesperiidae ()

Authors Zhang, J; Lees, David; Shen, J; Cong, Q; Huertas, B; Martin, G; Grishin, NV 195 ARTICLE The mitogenome of a Malagasy butterfly Malaza fastuosus (Mabille, 1884) recovered from the holotype collected over 140 years ago adds support for a new subfamily of Hesperiidae (Lepidoptera) Jing Zhang, David C. Lees, Jinhui Shen, Qian Cong, Blanca Huertas, Geoff Martin, and Nick V. Grishin

Abstract: Malaza fastuosus is a lavishly patterned butterfly from a that has three described species, all endemic to the mainland of Madagascar. To our knowledge, M. fastuosus has not been collected for nearly 50 years. To evaluate the power of our techniques to recover DNA, we used a single foreleg of an at least 140-year-old holotype specimen from the collection of the Natural History Museum London with no destruction of external morphology to extract DNA and assemble a complete mitogenome from next generation sequencing reads. The resulting 15 540 bp mitogenome contains 13 protein-coding genes, 22 transfer RNA genes, two ribosomal RNA genes, and an A+T rich region, similarly to other Lepidoptera mitogenomes. Here we provide the first mitogenome also for ( extrusus). Phylogenetic analysis of available skipper mitogenomes places Malaza outside of Trapezitinae and Barcinae + Hesperiinae, with a possible sister relationship to Heteropterinae. Of these, at least Heteropterinae, Trape- zitinae, and almost all Hesperiinae have monocot-feeding caterpillars. Malaza appears to be an evolutionarily highly distinct ancient lineage, morphologically with several unusual hesperiid features. The monotypic subfamily Malazinae Lees & Grishin subfam. nov. (type genus Malaza) is proposed to reflect this morphological and molecular evidence. Key words: next-generation sequencing, phylogeny, Madagascar, ancient DNA, Paul Mabille. Résumé : Le Malaza fastuosus est une espèce de papillon aux motifs très colorés qui appartient à un genre qui compte trois espèces décrites, toutes indigènes de Madagascar. Au meilleur de la connaissance des auteurs, aucun spécimen du M. fastuosus n’a été collecté depuis près de 50 ans. Pour évaluer la puissance des techniques d’extraction d’ADN, les For personal use only. auteurs ont employé une seule patte avant sur un spécimen holotype vieux d’au moins 140 ans, conservé au sein de la collection du Natural History Museum London et qui ne présentait aucun signe morphologique de dommage extérieur, pour extraire de l’ADN et assembler un mitogénome complet à l’aide de séquençage de seconde génération. Le mitogénome obtenu compte 15 540 pb et contient 13 gènes codant pour des protéines, 22 gènes d’ARN de transfert, deux gènes codant pour des ARN ribosomiques et une région riche en A+T, tout comme les autres mitogénomes de lépidoptères. Les auteurs rapportent également le premier mitogénome au sein des Trapezitinae (Rachelia extrusus). Une analyse phylogénétique parmi les autres mitogénomes d’hespéridés place le genre Malaza à l’extérieur des Trapezitinae et Barcinae + Hesperiinae, mais avec une possible relation proche avec les Heterop- terinae. Parmi celles-ci, les Heteropterinae, les Trapezitinae et presque tous les Hesperiinae présentent des che- nilles qui se nourrissent de monocotylédones. Le genre Malaza semble former un groupe distinct et ancien sur le plan évolutif, mais qui présente plusieurs caractéristiques morphologiques inhabituelles présentes chez les Hes- periidae. Une sous-famille monotypique, Malazinae Lees & Grishin (genre type Malaza) est proposé pour refléter Genome Downloaded from www.nrcresearchpress.com by Nick V. Grishin on 03/06/20 les évidences morphologiques et moléculaires. [Traduit par la Rédaction] Mots-clés : séquençage de seconde génération, phylogénie, Madagascar, ADN ancient, Paul Mabille.

Received 8 November 2019. Accepted 8 January 2020. J. Zhang* and J. Shen. Departments of Biophysics and Biochemistry, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390-8816, USA. D.C. Lees,* B. Huertas, and G. Martin. Department of Life Sciences, Natural History Museum, London, SW7 5BD, United Kingdom. Q. Cong. Institute for Protein Design and Department of Biochemistry, University of Washington, 1959 NE Pacific Street, HSB J-405, Seattle, WA 98195, USA. N.V. Grishin. Howard Hughes Medical Institute, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390-9050, USA; Departments of Biophysics and Biochemistry, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390-8816, USA. Corresponding author: Nick V. Grishin (email: [email protected]). *These authors contributed equally to this work. Copyright remains with the author(s) or their institution(s). Permission for reuse (free in most cases) can be obtained from RightsLink.

Genome 63: 195–202 (2020) dx.doi.org/10.1139/gen-2019-0189 Published at www.nrcresearchpress.com/gen on 6 March 2020. 196 Genome Vol. 63, 2020

Introduction round and glimmering white spangles tinted with yel- Its name translated from Malagasy as “famous”, Malaza low, whereas M. empyreus has three orange spots embed- Mabille, 1904 is probably the most unusual skipper genus ded in a similar vermilion-red. Malaza carmides is less from Madagascar. Three Malagasy endemic species of Malaza spectacular but sexually dimorphic, the female sporting have been described (Lees et al. 2003). One of them, aptly a single silver spangle on the ventral hindwing under- named Malaza fastuosus (Mabille, 1884), here dubbed Lavish side. Malaza, has not been collected to our knowledge for nearly Nothing is recorded about the habits of M. fastuosus, 50 years. The study specimen (Fig. 1; whose sex Paul but it has been collected widely, from north, west, east, and Mabille was unsure of: Mabille [1887]: 338), is a female, as southwest Madagascar (forêt d’Analalava, Ankarafantsika, ascertained by lack of androconial scale brushes on its Perinet, Ile Ste Marie, Fito, Forêt de Rogez, Imerimandroso ventral forewing and dorsal hindwing. Its taxonomic his- Lac Alaotra, and Tongobory; museum data and Viette (1956)), tory is somewhat complicated. It was first described by sites varying from 0 to 900 m elevation in dry to wet forests, Mabille in 1878 as his female syntype of Cyclopides empyreus from September to December. Surprisingly, it was last col- (Mabille 1878). Its type locality is no more precise than Mad- lected in Ankarafantsika in 1971, whereas M. empyreus is agascar and Mabille subsequently (re-)described this speci- occasionally collected or photographed in dense eastern men (Mabille 1884) as a distinct taxon, fastuosus. rainforest interiors (Preston-Mafham 1991). Malaza carmides The latter specimen (NHMUK010621353) is considered to is found more often in littoral forest or scrub. It is probable be the holotype of T. fastuosus, and apart from its head that additional species remain to be described. Searches for and abdomen, is consistent in pattern with the illustra- ovipositions for any species have so far proved fruitless, but tion in pl. 13, figs. 9, 9a of Mabille ([1885]) (those parts excepting M. carmides, Malaza encounters during fieldwork were probably fabricated in this figure). This holotype is are rare and not easily predictable. found in the collection of the Natural History Museum in The systematics of Malaza is poorly understood. Evans London (NHMUK). Even at the time of its description (1937) placed Malaza in the first subgroup of his Ploetzia as T. fastuosus, the specimen was “à corps incomplet” group, apparently largely based on wing venation “hind- (Mabille 1884) or indeed “privé d’abdomen et de tête” wing lower outer angle of cell not turned up, i.e., the (Mabille [1887]). As Mabille says in the last publication that it median vein and vein 4 collinear” (an observation appar- was lacking its head and abdomen, it may be inferred very ently not based on a wing preparation of Malaza, see likely that the abdomen, currently attached with dark red Diagnosis). This group he considered contained also resin, is not the original one (it lacks an expected series of dark Miraja and Perrotia, two other genera endemic to Mada- ventro-lateral stripes, and is slightly larger than normal for gascar (see also Lees et al. 2003, where Miraja was sunk). Malaza carmides). The current wingspan measures just short An additional 12 Afrotropical genera were included in For personal use only. of 49 mm, but the tip of the left forewing is missing today. Evans' Ploetzia group. Viette (1956: 46–50) added a limited Mabille measured its wingspan at 52 mm in 1878 and morphological study on the genus, illustrating the male 53 mm in 1884, both measurements being consistent with and female genitalia of two species. Although these skip- the specimen now numbered NHMUK010621353 being both pers are robust and striking, no authors have proposed the supposed female syntype of C. empyreus and the holo- any fundamentally outstanding morphological charac- type of T. fastuosus. Once Mabille acquired a true female of ters of Malaza in a comparative context, although Mabille Malaza empyreus, of similar size to its male, he apparently (1904) mentioned the inflated hindleg first tarsal seg- realized his mistake (Mabille [1887]: 337–338). While it is ment peculiar to M. carmides. Although the genus has not not known when and exactly where or by whom the spec- previously been included in molecular analyses, two spe- imen was collected, the collection event was prior to cies have been DNA barcoded by D.C.L. (M. carmides,

Genome Downloaded from www.nrcresearchpress.com by Nick V. Grishin on 03/06/20 15 December 1878, the dating of the page where the descrip- M. empyreus) but neighbor-joining and ad hoc phylogenetic tion of the female syntype of C. empyreus was published analyses of barcodes fails to place the genus reliably. (Mabille 1878: 285), alongside other species typical of east- To learn more about these Malagasy skippers, we se- ern lowland rainforest described there, notably M. carmides quenced, assembled, and annotated the complete mitog- (as Cyclopides catocalinus), Xanthodisca ariel, and Perrotia gillias. enome of the Malaza fastuosus holotype. The foreleg was Malaza fastuosus is the largest and most sumptuously used for DNA work, because it remains to be verified if the colored of the three species of Malaza and among the most abdomen currently affixed to NHMUK010621353 belongs ei- stunning skippers in the Afrotropical region. The genus ther to this species or even to Malaza. Methods for genomic exhibits a rather stout, thorax and abdomen and both sexes DNA extraction, library construction, next-generation se- are brown above with two orange bands on the hindwing quencing, and computational procedures have been re- and orange hindwing fringes. Ventrally, however, M. fastuosus ported by us previously (Cong and Grishin 2016; Shen resembles only the much smaller M. empyreus, the type spe- et al. 2015, 2016). Importantly, after being detached, the cies of Manarina Mabille, 1904 (Lucas 1905: 699). Richly col- leg was incubated in the DNA extraction buffer as a ored in various shades of vermilion and brown, each whole, without being broken. This procedure allows for hindwing of M. fastuosus is endowed with four large the preservation of body parts of unique specimens and

Published by NRC Research Press Zhang et al. 197

Fig. 1. Study specimen and phylogenetic tree. Malaza fastuosus holotype, female (specimen NHMUK010621353, molecular code 0247278458, sample NVG-18082C03) with its labels is shown on the left: dorsal view (above), ventral view (middle), and labels (below). Scale bar refers to all images. On the right, is the maximum likelihood tree of the coding regions in the mitogenomes of 26 Hesperiidae species. The tree is rooted with Pterourus glaucus (Papilionidae). MrBayes posterior probabilities are shown by each node. Species names for mitogenomes sequenced in this study are shown in color. GenBank accessions for mitogenome sequences are as follows: NC_030602.1; Agathymus mariae KY630504.1; Apostictopterus fuliginosus NC_039946.1; Barca bicolor NC_039947.1; Burara striata NC_034676.1; Carterocephalus silvicola NC_024646.1; Celaenorrhinus maculosa NC_022853.1; Choaspes benjaminii NC_024647.1; Ctenoptilum vasava NC_016704.1; Daimio tethys NC_024648.1; Erynnis montanus NC_021427.1; Euschemon rafflesia NC_034231.1; Hasora anura NC_027263.1; Hasora vitta NC_027170.1; Heteropterus morpheus NC_028506.1; NC_029826.1; Lobocla bifasciatus NC_024649.1; Malaza carmides MN919191; Malaza empyreus MN919190; Malaza fastuosus MK301537; Megathymus yuccae KY630500.1; Parnara guttatus NC_029136.1; Potanthus flavus NC_024650.1; Pterourus glaucus NC_027252; Pyrgus maculatus NC_030192.1; Rachelia extrusus MN919192; Tagiades vajuna KX865091.1. For personal use only.

does not impair subsequent DNA work. From this leg, we read length after trimming adapters reaches a maxi- extracted 0.78 ng of genomic DNA (gDNA). Due to the low mum, usually about 150 bp. The mitogenome of Lerema

Genome Downloaded from www.nrcresearchpress.com by Nick V. Grishin on 03/06/20 amount of gDNA, we did not check DNA size using gel, accius (Smith, 1797) (Cong and Grishin 2016) was used as a but proceeded with pair-end DNA library construction reference to search for (“bait”) similar sequence reads for next generation sequencing. The final library was using BWA (Li and Durbin 2009). Nearly 0.95% (520 325 7.1 ng/␮L, 40 ␮L volume, with the average size of about out of 54 557 740) of M. fastuosus total genomic reads were 190 bp. Since sequencing adapters are about 125 bp, the extracted by BWA for mitogenome assembly (Hahn et al. library size of 190 bp suggests an average genomic DNA 2013). Despite the DNA being heavily fragmented, fragment of about 65 bp. Sequencing of the library on an a number of reads were longer than 100 bp, allowing us Illumina HiSeq X10 instrument for 150 bp at both ends to confidently assemble the mitogenome. The complete resulted in 55.6M raw reads and, after trimming adapt- mitogenome of M. fastuosus was assembled de novo using ers, 53.5M reads with lengths above 30 bp were left, to- Platanus (Kajitani et al. 2014) followed by a manual gap- taling about 3 Gbp. Overall, more than 92% of reads had closing procedure. lengths from 20 to 70 bp, with the median around 47 bp, The mitochondrial genome of M. fastuosus is 15 540 bp suggesting highly degraded genomic DNA. For samples in length (GenBank: MK301537) and is AT rich, with a base field-preserved in a DNA-friendly way, we typically ob- composition of 40.1% A, 40.4% T, 7.4% G, and 12% C. It en- tain from 100 ng to several milligrams of gDNA, and the codes a typical gene set for Lepidoptera mitogenomes,

Published by NRC Research Press 198 Genome Vol. 63, 2020

including 13 protein-coding genes (ND1–ND6, COX1–COX3, coding sequences were further partitioned by 1st codon ND4L, ATP8, ATP6, and CYTB), 22 tRNA genes (two for (pos1), 2nd codon (pos2), and 3rd codon (pos3). We used serine and leucine and one for each of the rest of the the partitioning scheme that was developed for Hesperii- amino acids), two ribosomal RNAs (rrnL and rrnS), and an dae mitogenomes in a dedicated study (Han et al. 2018). A+T rich D-loop control region. The typical start codon Accordingly, sequences were partitioned into 16 subsets. ATN is used in other genes except COX1, which has a GTG Subset 1: tRNA-Phe, tRNA-His, ND6-pos1, ND2-pos1, tRNA- codon. The stop codon in COX2, ND4, and ND5 genes is Leu, ATP6-pos1, and ND3-pos1; subset 2: tRNA-Lys, tRNA- incomplete (just T), and a complete TAA codon should be Leu, ATP6-pos2, COX2-pos2, COX3-pos2, and CYTB-pos2; formed during mRNA maturation (Boore 1999; Ojala subset 3: ATP6-pos3 and ND2-pos3; subset 4: ATP8-pos1 et al. 1981). The two rRNAs are 1479 bp (rrnL) and 773 bp and ATP8-pos2; subset 5: ATP8-pos3, ND6-pos3, COX2- (rrnS) and the tRNAs are from 66 to 72 bp in length. pos3, CYTB-pos3, COX1-pos3, COX3-pos3, and ND3-pos3; Previously, by Sanger sequencing we had obtained subset 6: COX1-pos1; subset 7: COX1-pos2; subset 8: tRNA- DNA barcodes (COX1-5P) for the two other species of Gln and COX2-pos1; subset 9: COX3-pos1, CYTB-pos1, Malaza: M. empyreus and M. carmides (on BOLD: cluster num- tRNA-Trp, and tRNA-Met; subset 10: ND4-pos2, tRNA-Ile, bers BOLD:AAK7349, BOLD:ACW7233, respectively, based ND4L-pos1, ND1-pos1, and ND5-pos2; subset 11: ND4L- on single specimens that D.C.L. collected in 2004 from pos2, ND5-pos3, and ND1-pos2; subset 12: ND1-pos3; sub- Ambatovy, east Madagascar). Comparison with the bar- set 13: ND3-pos2, ND6-pos2, and ND2-pos2; subset 14: code of M. fastuosus reveals that M. empyreus is more sim- ND4-pos3 and ND4-pos1; subset 15: ND5-pos1 and ND4L- ilar than is M. carmides: 92.7% sequence identity (6.3% pos3; subset 16: tRNA-Pro, tRNA-Tyr, tRNA-Val, 12S ribo- pairwise divergence), in accord with morphological sim- somal RNA, 16S ribosomal RNA, tRNA-Asp, tRNA-Glu, ilarities (e.g., wing pattern, limited sexual dimorphism), tRNA-Arg, tRNA-Asn, tRNA-Gly, two tRNA-Ser, tRNA-Ala, as well as apparent phylogenetic closeness. Malaza fastuosus and tRNA-Thr. The best substitution rate models for each and M. empyreus were placed in the same genus Manarina partition were selected by MrBayes. by Mabille (1904: 95–96). The barcode of M. carmides is The resulting tree topology is consistent with previous more distant at around 90.5% sequence identity (9.4% phylogenetic studies that did not include Malaza (Sahoo pairwise divergence between M. carmides and M. fastuosus; et al. 2016; Warren et al. 2008, 2009; Zhang et al. 2017a, 9.6% between M. carmides and M. empyreus). The differ- 2019). Coeliadinae are the sister to all other Hesperiidae; ence of less than 10% pairwise divergence for the three Euschemon is a sister to the rest of Hesperiidae except species of Malaza and strong similarities in wing vena- Coeliadinae; the relationship between and tion (see below) suggests that they are close phylogenetic Pyrginae is unresolved (0.79 posterior probability) but relatives and could be congeners. This small difference with Pyrginae sister to Tagiadinae, and Heteropterinae For personal use only. also indirectly validates our sequencing result, suggest- are in the same clade with Hesperiinae (Zhang et al. 2019) ing that the sequence we obtained is indeed of a third spe- (the last two subfamilies are monocot feeders as caterpil- cies of Malaza, and not some contaminant. This is an lars, with secondary colonization of dicots by some Hes- essential point, since above we mentioned the specimen periinae). Malaza is sister to Heteropterinae with posterior sequenced might be chimaeric. However, when the foreleg probability of 0.99. This relationship comes as a surprise was detached by N.V.G., there was no evidence that it had and has not hithero been suggested by morphology. been glued there. Therefore, additional studies, in particular those based To better understand the phylogenetic affinities of on the analysis of nuclear genomes, are needed to firmly Malaza, we additionally assembled mitogenomes of establish phylogenetic affinities of Malaza. However, M. carmides and M. empyreus from the same vouchers used even with the data at hand, it is clear that Malaza origi-

Genome Downloaded from www.nrcresearchpress.com by Nick V. Grishin on 03/06/20 for DNA barcoding using the same procedures as for M. nated early during the radiation of monocot feeders and fastuosus. Furthermore, because no mitogenomes from represents an ancient lineage, today surviving only in the subfamily Trapezitinae were available, we assembled Madagascar. It is noteworthy that despite some pheno- a mitogenome of Rachelia extrusus (C. & R. Felder, 1867) typic similarities with Malaza (e.g., bulky, robust bodies) using the same approach. A Bayesian inference tree was that may have led to Mabille’s (1884) placement in constructed by MrBayes (Huelsenbeck and Ronquist Trapezites, Trapezitinae (represented by Rachelia extrusus) 2001) from representative known Hesperiidae mitog- is not closely related to it and is instead in the same clade enomes (Cao et al. 2016; Cong and Grishin 2016; Han et al. with Barcinae and Hesperiinae (with 0.98 posterior prob- 2018; Hao et al. 2012; Kim et al. 2014; Liu et al. 2017; Shao ability support; Barcinae + Trapezitinae with marginal et al. 2015; Shen et al. 2015, 2016; Wang et al. 2014; Wang support at 0.95). et al. 2015; Wang et al. 2016; Zhang et al. 2017a, 2017b) To probe its morphology further, D.C.L. examined with our four new mitogenomes added, and rooted with specimens and preparations of Malaza, including genita- the Pterourus glaucus (Papilionidae) mitogenome (Shen et al. lia dissections at NHMUK (H. 828, H. 901, H. 902, BMNH(E) 2015)(Fig. 1). We used all protein coding regions, tRNAs 526064 - dry genitalia on card) and wing venation for (except tRNA-Cys), and two ribosomal RNAs. Protein- M. carmides and M. empyreus, for characters of potential

Published by NRC Research Press Zhang et al. 199

relevance in the systematics of Hesperiidae. Reinforced dae specimens from all major phylogenetic groups in- by the mitogenomic tree, and its morphological distinct- cluded in Zhang et al. (2019). These base pairs are conserved ness from the other three subfamilies, we conclude that in all three species of Malaza and uniquely place Malaza the Malaza lineage should have the same taxonomic rank within the monocot-feeding Hesperiidae to diagnose as Heteropterinae, Trapezitinae, Barcinae, and Hesperii- Malazinae as distinct from the three other subfamilies, nae (see Zhang et al. 2019). To reflect the ancient origin of i.e., base pairs that are (i) the same as in Heteropterinae Malaza and its distinctness from other Hesperiidae, a new but different from most Hesperiinae and Trapezitinae subfamily for it is proposed. (604); or (ii) the same as in most Hesperiinae but different from most Heteropterinae and Trapezitinae (319, 481); or Malazinae Lees and Grishin, subfam. nov. (iii) the same as in Trapezitinae, but different from most ZOOBANK NO.: http://zoobank.org/E4D5B92E-8E4A-4BEC-812B- Hesperiinae and Heteropterinae (475); or (iv) the same as 5B1AA226FC0E. most Trapezitinae and Heteropterinae but mostly differ- TYPE GENUS: Malaza Mabille, 1904, by original designation. ent from Hesperiinae (38, 512); or (v) the same as in few or most Hesperiinae and Trapezitinae, but different from DIAGNOSIS: Morphologically, distinguished from other Hes- periidae or other subfamilies, by the following characters Heteropterinae (451, 475, 479). None of these characters or their combination (additional contrasts given in de- is a unique synapomorphy of Malaza, but it may be noted scription): Head. 1. Compound eyes covered in short in- that 38 is a first position change coding for valine (whereas terommatidial setae, a very unusual feature in hesperiids, leucine in Hesperiinae) and 512 is a first position change recorded in some coeliadines, notably Choaspes and a few coding for alanine (but serine in Hesperiinae). Hasora, not known elsewhere (Warren et al. 2009) and also DESCRIPTION: Generalities. Medium to large hesperiids, known in Macrosoma (Yack et al. 2007). 2. A short but forewing length 12.5–24 mm, wingspan (apex–apex) 30– prominent eyelash present, a synapomorphy for non- 50 mm, wings broad and rounded, slightly narrower and coeliadine hesperiids. 3. Posterior and anterior cephalic more acute in male, very sexually dimorphic or little so, chaetosemata present, again a synapomorphy for non- abdomen similar length to hindwings, body very robust, coeliadine hesperiids, anterior pair concentrated into head relatively broad. Wings held over back, downwards, at lateral bunches of radially oriented fine setae just ante- rest (basking posture not observed). Proboscis quite long, riad of eyelash, posterior pair linearly radiating and al- flower visiting, diurnal, life history unknown, but expressed egg of M. carmides is quite large (around 1 mm diameter) most meeting at midline. Thorax. 4. Forewing M2 weak, and domed, hemispherical, orange, with around 17 radial curving slightly towards M3 near and at cell; M2 originat- ribs. Head. Compound eyes covered in short interomma- ing distinctly closer to M3 than M1 as in most Hesperii- tidial setae. Eyelash and posterior and anterior cephalic For personal use only. nae, but not so extreme as in most hesperiine cases (at chaetosemata present (see diagnosis). Labial palps erect, cell, M1-M2:M2-M3 about 0.55:0.45: respectively); M2 ori- gin is not intermediate (Kitching et al. 1999) nor tending partly straight, second segment irregularly bulbous but relatively flattened (quadrate would not be a precise de- towards M1 as in some Trapezitinae (Braby 2000: 88, al- though the opposite case is indicated on p. 22). 5. Hind- scription (Evans 1937; Viette 1956)) rather than porrect and hairy as in Heteropterinae, collapsed inwards in wings with M2 lost; discocellulars (cross-veins) also lost their proximal part towards head, third segment in line except for stumps visible after fork between M1-Rs and with second segment (i.e., erect; observed also in natural M3-CuA1, respectively. Extrapolated chord between stumps and their alignment with tornus also more or less perpen- position), relatively short and stout (about twice as long dicular to costa and upper cross-vein [remnant] certainly as maximum width), slightly fatter than conical, pointed not directed towards the apex, as has been supposed to be towards tip. Antennae about 1/2 costal length, as, e.g., in

Genome Downloaded from www.nrcresearchpress.com by Nick V. Grishin on 03/06/20 a synapomorphy of Trapezitinae, nor for the lower part trapezitines (not shorter than mid-costa, as in many Het-

of the M2-M3 discocellular, directed towards the body eropterinae), strongly downbent well after commencement of (Braby 2000; Simonsen et al. 2012; Warren et al. 2009). 6. club (strongly hooked), long apiculus tapered to pointed tip. Hindwing cell (or trace of it up to cross-vein stumps), Thorax. Foreleg epiphysis present, well developed, not re- about half wing length, not long (as typical of Heterop- duced as suggested for most heteropterines (Warren et al. terinae) or fairly long (as in Trapezitinae). The hairy com- 2009). Mesothoracic tibia with one pair of spines and

pound eye and loss of a tubular hindwing M2 and metathoracic tibia with one pair of spines (in M. carmides, associated discocellulars are the clearest presently M. fastuosus;(Mabille 1904)) or two, as in some Trapezitinae known synapomorphies for Malazinae although proba- (Kitching et al. 1999: 83) with mid tibial pair shorter bly they are also parallelisms exhibited elsewhere in Hes- (M. empyreus). Hindleg first tarsal segment strongly in- periidae. flated (not the tibia as misstated by Evans 1937: 132) in By DNA, a combination of the following base pairs in male of M. carmides (Mabille 1904: 95), excavate length- COI barcode is diagnostic: 38G, 319A, 451T, 475T, 479C, wise towards the body like a hollowed out boat, not itself 481T, 512G, and 604A. To derive these DNA characters, we clothed internally with obvious androconial scales but used a barcode alignment of 160 representative Hesperii- harboring the dark tip of a brush originating along the

Published by NRC Research Press 200 Genome Vol. 63, 2020

length of the femur, bundled with a larger recumbent stumps (see diagnosis). No trace of median fork before hair pencil originating at the proximal end of the tibia; cell end (as in trapezitines and a few other taxa, see otherwise, hindtibiae unmodified except fringed with Parsons 1999: 152) and, where stumps may be substan- orange hairs. The presence in at least one Malazinae of a tial, top cross-vein stump oriented towards costa rather recumbent metatibial hair tuft (common in Coeliadinae) than towards body or apex. Extrapolated chord between stands in contrast to its apparent absence all three prior stumps and their alignment also more or less perpendic- monocot feeding subfamilies (Warren et al. 2009). ular to costa in line with tornus (although inaccurate to Wings. Forewing with costal yellow-orange streak in state: end of cell straight (Viette 1956)); and upper cross- basal half on both surfaces, except in M. carmides.In vein [remnant] certainly not directed towards the apex M. empyreus and M. fastuosus, hindwing dorsum with or- (or termination of Rs), as has been supposed to be a sy- ange spot at wing base and end of cell the latter a con- napomorphy of Trapezitinae (Simonsen et al. 2012; tinuation of orange crescent at base of cubitus, two-three Warren et al. 2009). Hindwing cell (or trace of it up to cross-vein stumps), about half wing length, not long (as orange spots more distally in space CuA2, with two extra typical of Heteropterinae). Rs fork with stem originating orange flecks in space M3 in M. carmides making a linear series of four. Hindwing ventrally with pale spangles in well before fork of CuA2 with its stem (Viette 1956). Sex scales. See hindleg of M. carmides, above. Ventral fore- similar positions, two along space CuA2 and 1–2 either end of cell, as silver spangles in M. fastuosus or as similarly wing brush of long fine pale bristles towards base of shaped (but larger) orange marks in M. empyreus, which space CuA1 (M. empyreus, M. fastuosus)(Evans 1937: 132; lacks the silvery basal cell spangle present in M. fastuosus. Viette 1956). Also on dorsal hindwing of male, extensive These marks are set in vermilion-red areas to dark or- (dark in M. empyreus or light yellowish in M. fastuosus, ange red areas, the latter much more extensive in the absent in M. carmides) brush of scales at base of (originat- hindwing in M. empyreus, which has more defined black ing mainly above) Sc+R1 and another brush, less dense, at spots, three along the cell region and two along space base of cell (black in males of M. fastuosus and M. empyreus, CuA , and submarginal black spots in the spaces of the yellowish in both sexes in M. carmides), interspersed with 2 lenticular patch of minute grey relatively shiny scales ventral hindwing not conspicuous in the other species. (latter in all three species). Patches of fine thin pale scales Hindwing fold, widely channeled ventrally, in male ter- in ventral forewing at base of space 1A+2A and at base of gad of 1A+2A bearing dorsally an extensive tuft of yellow- space CuA (M. empyreus, M. carmides). Abdomen. As long as ish scales in both sexes. Malaza fastuosus has particularly 1 hindwings, in M. empyreus and M. fastuosus, underside or- pronounced orange cilia along the hindwing margin, ange flanked with black streaks, in M. carmides, plain other species with narrower fringe. Patterning similar in orange. Male genitalia. Uncus much shorter than tegu- For personal use only. sexes of M. empyreus and M. fastuosus, whereas striking men, appearing slightly hooked at tip from lateral view, sexual dimorphism in M. carmides, the female’s hindwing spatulate-truncate from dorsal view and marginally bifid ventral surface with a single large silver spangle only in (in M. carmides) to distinctly bifid (in M. fastuosus and M. the cell, matched by an orange crescent dorsally, and an empyreus), similar to some trapezitines (Kitching et al. arc of four orange arrowheads in space M -CuA . Unlike 3 2 1999: 82; Viette 1956: fig. 51). Gnathos large to very large in most Trapezitinae and Hesperiinae where they tend to (M. fastuosus preparation H. 828, #18516), sclerite some- be well developed, hyaline patches in forewing lost, or they what rectangular from lateral view but expanded later- are reduced in the case of the male of M. carmides, except in ally (in all three species; see also fig. 51 in Viette (1956), the female where a single hyaline patch in the forewing cell for M. carmides, gnathos apparently incomplete). Valves and another two in space M3 and space CuA1, are clearly at the apex in the costal region and the distal region of the defined on both surfaces as quadrate markings (some- Genome Downloaded from www.nrcresearchpress.com by Nick V. Grishin on 03/06/20 sacculus (harpe) darkly sclerotized, with several short times also weakly expressed in the female of M. empyreus). teeth. In M. carmides, the costal and distal lobes are In the ventral forewing of M. carmides, there is an extensive rounded and the distal lobe of the sacculus bears about medial pale area diffusely in spaces CuA2-1A+2A. Apart 6–8 teeth, whereas in M. empyreus and M. fastuosus the from this and the hyaline traces, the ventral surface of main valve is squared off apically with a series of about the male of M. carmides is more or less unmarked, except 20–30 much finer teeth with further smaller ones to for a pale trace of the large hindwing cell spot on the their side and the more rounded sacculus lobe bears a dorsum. The ventral surface of M. carmides has a violet smaller number of rather finer teeth (Viette 1956: figs. 51, sheen missing in the other species. Wing venation. 52). Saccus very short. Aedeagus straight, without unusual Forewings with CuA2 originating relatively close to wing modifications. Female genitalia. Hairy papillae anales

base (primary fork with stem of M and CuA1 at about 1/8 variably shaped, bearing a pair of long apophyses poste- of total length), as, e.g., in some coeliadines (e.g., Bibasis riores that extend dorsally well beyond the shield-like oedipodia (Swainson, 1820) (Bascombe et al. 1999: 76)). postvaginal median lobe which has a medially indented

Forewing M2 as for diagnosis. Hindwings with M2 lost slot distad, as far as a pair of narrow ventro-lateral scle- and cell end cross-veins (discocellular veins) reduced to rites (shaped like curly braces) bearing flaps mediad. In

Published by NRC Research Press Zhang et al. 201

between these sclerites ventrally is a median antevaginal Data deposition plate in M. carmides, absent in M. empyreus and M. fastuosus. The mitogenomes reported in this study were deposited Further proximad are paired sclerotized lobes (absent in in GenBank with the accessions MK301537, MN919190– M. carmides) densely covered in long setae that continue to MN919192. This publication has been registered with the intersegmental membrane of A6/A7 imparting a ZooBank as http://zoobank.org/8E2333AE-FEBC-4025-AA02- “silky” appearance (in M. fastuosus and M. empyreus) DE4BFC28E3F7 and the taxonomic act (new subfamily) was registered as http://zoobank.org/E4D5B92E-8E4A- (Viette 1956: fig. 53). Corpus bursae linked by a more or 4BEC-812B-5B1AA226FC0E. less narrow ductus bursae. Ostium bursae of female in all species with extensive paired spiculose signa consisting Disclosure statement of two narrow lateral weakly sclerotized areas bearing The authors declare no conflicts of interest. numerous minute triangular spines on internal walls Funding information leading up to a sclerotized “neck” distad before the en- This work was supported in part by the grants to trance of the ductus bursae. Warren et al. (2009) did not N.V.G. from the National Institutes of Health (GM094575 score this character, so its distribution among Hesperii- and GM127390) and the Welch Foundation (I-1505). D.C.L. dae subfamilies is not fully clear, but it occurs in some acknowledges Leverhulme Trust (F/00696/I), for fieldwork Heteropterinae (Steinhauser 1991). Appendix bursae funding, ERC-EMARES (#250325), and Malagasy authorities clearly present and non-spiculose (at least in M. fastuosus including MICET for assistance in logistics and permits preparation H.901, #18568; apparently also present in (No. 0 – MINEV.EF/SG/DGEF/DPB/SCBLF). M. empyreus, but available slide preparation H.902, #18569 was not adequate to confidently score this character) nor References Bascombe, M.J., Johnston, G., and Bascombe, F.S. 1999. Butter- was it clear from constrictions indicated in diagrams of flies of Hong Kong. Academic Press, London, United King- M. fastuosus and M. carmides, figs. 53, 54 in Viette (1956), dom. but this is a structure well developed in Heteropterinae Boore, J.L. 1999. mitochondrial genomes. Nucleic Acids and Trapezitinae in contrast to Hesperiinae, with a few Res. 27(8): 1767–1780. doi:10.1093/nar/27.8.1767. PMID:10101183. exceptions (e.g., Ceratrichia clara and Meza meza: Warren Braby, M.F. 2000. The butterflies of Australia: their identifica- tion, biology and distribution. CSIRO Publishing, Melbourne. et al. 2009). Cao, L., Wang, J., James John, Y., Yau Shiu, M., Yuan, X., Liu, J., GENERA INCLUDED: Malaza and its junior subjective syn- and Cao, T. 2016. The complete mitochondrial genome of onym Manarina Mabille, 1904. Hasora vitta (Butler, 1870) (Lepidoptera: Hesperiidae). Mito- chondrial DNA A DNA Mapp. Seq. Anal. 27(4): 3020–3021. The Malazinae is the third monotypic subfamily of doi:10.3109/19401736.2015.1063048. PMID:26152350.

For personal use only. Hesperiidae after the Indian Chamundinae (Zhang et al. Cong, Q., and Grishin, N.V. 2016. The complete mitochondrial 2019) and the Euschemoninae. The latter consists of a genome of Lerema accius and its phylogenetic implications. single species, Euschemon rafflesia (MacLeay, [1826]). It is an PeerJ, 4: e1546. doi:10.7717/peerj.1546. PMID:26788426. de Jong, R. 2007. Estimating time and space in the evolution of Australian endemic and the only butterfly with a frenu- the Lepidoptera. Tijdschrift voor Entomologie, 150: 319–346. lum and retinaculum (apart from all but three species of doi:10.1163/22119434-900000233. Hedylidae), in males only in both families (Braby 2000; Evans, W.H. 1937. A catalogue of the African Hesperiidae indi- Scoble 1986), structures commonly present in moths to cating the classification and nomenclature adopted in the British Museum. The Trustees of the BMNH, London, United couple the wings. Euschemon rafflesia is also a large showy Kingdom. skipper with black, yellow- or white-marked wings and a Hahn, C., Bachmann, L., and Chevreux, B. 2013. Reconstructing red abdomen tip. It is interesting that these ancient mitochondrial genomes directly from genomic next-generation monotypic lineages survive in locations characterised by sequencing reads–a baiting and iterative mapping approach.

Genome Downloaded from www.nrcresearchpress.com by Nick V. Grishin on 03/06/20 Nucleic Acids Res. 41(13): e129. doi:10.1093/nar/gkt371. PMID: high endemism (India, Australia, and Madagascar). This 23661685. reinforces the idea that Hesperiidae as a family might have Han, Y., Huang, Z., Tang, J., Chiba, H., and Fan, X. 2018. The originated and diversified around the Indian Ocean (while the complete mitochondrial genomes of two skipper genera sister family Hedylidae is neotropical only). The earliest di- (Lepidoptera: Hesperiidae) and their associated phylogenetic analysis. Sci. Rep. 8(1): 15762. doi:10.1038/s41598-018-34107-1. verging hesperiid subfamily, Coeliadinae, is also palaeotropi- PMID:30361496. cal and suggested to be of possible oriental origin (de Jong Hao, J., Sun, Q., Zhao, H., Sun, X., Gai, Y., and Yang, Q. 2012. The 2007), but Madagascar actually contains the earliest diverg- complete mitochondrial genome of Ctenoptilum vasava (Lepi- ing member in one of the two clades in this subfamily, doptera: Hesperiidae: Pyrginae) and its phylogenetic implica- tion. Comp. Funct. Genomics, 2012: 328049. doi:10.1155/2012/ Tekliades ramanatek (Boisduval, 1833) (Li et al. 2019: S1 Appen- 328049. PMID:22577351. dix, p. 91). Therefore, Hesperiidae as a family might have Huelsenbeck, J.P., and Ronquist, F. 2001. MRBAYES: Bayesian originated and diversified around the Indian Ocean. Malazi- inference of phylogenetic trees. Bioinformatics, 17(8): 754– nae is the third supratribal endemic lineage of Lepidoptera 755. doi:10.1093/bioinformatics/17.8.754. PMID:11524383. Kajitani, R., Toshimoto, K., Noguchi, H., Toyoda, A., Ogura, Y., so far identified for Madagascar (after Whalleyanidae and Okuno, M., et al. 2014. Efficient de novo assembly of highly Callidulidae: Griveaudiinae). heterozygous genomes from whole-genome shotgun short

Published by NRC Research Press 202 Genome Vol. 63, 2020

reads. Genome Res. 24(8): 1384–1395. doi:10.1101/gr.170720. Shen, J., Cong, Q., and Grishin, N.V. 2015. The complete mito- 113. PMID:24755901. chondrial genome of Papilio glaucus and its phylogenetic im- Kim, M.J., Wang, A.R., Park, J.S., and Kim, I. 2014. Complete plications. Meta Gene, 5: 68–83. doi:10.1016/j.mgene.2015.05. mitochondrial genomes of five skippers (Lepidoptera: Hes- 002. PMID:26106582. periidae) and phylogenetic reconstruction of Lepidoptera. Shen, J., Cong, Q., and Grishin, N.V. 2016. The complete mitog- Gene, 549(1): 97–112. doi:10.1016/j.gene.2014.07.052. PMID: enome of Achalarus lyciades (Lepidoptera: Hesperiidae). Mito- 25058696. chondrial DNA B Resour. 1(1): 581–583. doi:10.1080/23802359. Kitching, R.L., Scheermeyer, E., Jones, R.E., and Pierce, N.E. 2016.1197070. PMID:28367503. 1999. Biology of Australian butterflies. CSIRO Publishing, Simonsen, T.J., de Jong, R., Heikkilä, M., and Kaila, L. 2012. But- Melbourne. terfly morphology in a molecular age. Does it still matter in Lees, D.C., Kremen, C., and Raharitsimba, H. 2003. Classifica- butterfly systematics? . Struct. Dev. 41: 307–322. tion, diversity and endemism of the butterflies (Papilion- doi:10.1016/j.asd.2012.04.006. PMID:22583793. oidea and Hesperioidea): a revised species checklist. In The Steinhauser, S.R. 1991. Taxonomic notes and descriptions of natural history of Madagascar. Edited by S.M. Goodman and new taxa in the Neotropical Hesperiidae. Part II, Heteropteri- J.P. Benstead. University of Chicago Press, Chicago, Ill. nae and Hesperiinae, Vinius group. Bull. Allyn Mus. 132: 1–47. pp. 762–793. Viette, P. 1956. Insectes: Lépidoptères Hesperiidae. In Faune de Li, H., and Durbin, R. 2009. Fast and accurate short read Madagascar. IRSM Tananarive, Tsimbazaza, Madagascar. alignment with Burrows-Wheeler transform. Bioinformat- pp. 1–85. ics, 25(14): 1754–1760. doi:10.1093/bioinformatics/btp324. Wang, A.R., Jeong, H.C., Han, Y.S., and Kim, I. 2014. The complete PMID:19451168. mitochondrial genome of the mountainous duskywing, Erynnis Li, W., Cong, Q., Shen, J., Zhang, J., Hallwachs, W., Janzen, D.H., montanus (Lepidoptera: Hesperiidae): a new gene arrangement and Grishin, N.V. 2019. Genomes of skipper butterflies reveal in Lepidoptera. Mitochondrial DNA, 25(2): 93–94. doi:10.3109/ extensive convergence of wing patterns. Proc. Natl. Acad. Sci. 19401736.2013.784752. PMID:23586336. USA. 116(13): 6232–6237. doi:10.1073/pnas.1821304116. PMID: Wang, J., James John, Y., Xuan, S., Cao, T., and Yuan, X. 2016. The 30877254. complete mitochondrial genome of the butterfly Hasora Liu, F.-F., Li, Y.-P., Jakovlic, I., and Yuan, X.-Q. 2017. Tandem anura (Lepidoptera: Hesperiidae). Mitochondrial DNA Part A, duplication of two tRNA genes in the mitochondrial genome 27(6): 4401–4402. doi:10.3109/19401736.2015.1089543. PMID: of Tagiades vajuna (Lepidoptera: Hesperiidae). Eur. J. Entomol. 26466192. 114(1): 407–415. doi:10.14411/eje.2017.052. Wang, K., Hao, J., and Zhao, H. 2015. Characterization of complete Lucas, R. 1905. Lepidoptera für 1904. Archiv für Naturgeschichte Jahrg.71: Bd.2: Heft.2: pt.1: 515–786. mitochondrial genome of the skipper butterfly, Celaenorrhinus Mabille, P. 1878. Diagnoses de Lépidoptères de Madagascar. Pe- maculosus (Lepidoptera: Hesperiidae). Mitochondrial DNA, 26(5): tites Nouvelles Entomologiques, 2(10): 285. 690–691. doi:10.3109/19401736.2013.840610. PMID:24102602. Mabille, P. 1884. Descriptions de Lepidopteres Exotiques. Annales Warren, A.D., Ogawa, J.R., and Brower, A.V.Z. 2008. Phyloge- de la Société Entomologique de Belgique, 28: CLXXXIV–CXCI. netic relationships of subfamilies and circumscription of Mabille, P. [1885]. Histoire naturelle des Lépidoptères. In His- tribes in the family Hesperiidae (Lepidoptera: Hesperioidea). toire physique naturelle et politique du Madagascar. Edited by Cladistics, 24(5): 642–676. doi:10.1111/j.1096-0031.2008. 00218.x. For personal use only. A. Grandidier. Imprimerie Nationale, Paris. 55 Pls. Mabille, P. [1887]. Histoire naturelle des Lépidoptères. In His- Warren, A.D., Ogawa, J.R., and Brower, A.V.Z. 2009. Revised toire physique naturelle et politique du Madagascar. Edited by classification of the family Hesperiidae (Lepidoptera: Hespe- A. Grandidier. Imprimerie Nationale, Paris. 364 pp. rioidea) based on combined molecular and morphological Mabille, P. 1904. Lepidoptera Rhopalocera. Fam. Hesperidae. In data. Syst. Entomol. 34(3): 467–523. doi:10.1111/j.1365-3113. Genera Insectorum. Edited by P. Wytsman. 2008.00463.x. Ojala, D., Montoya, J., and Attardi, G. 1981. tRNA punctuation Yack, J.E., Johnson, S.E., Brown, S.G., and Warrant, E.J. 2007. The model of RNA processing in human mitochondria. Nature, eyes of Macrosoma sp. (Lepidoptera: Hedyloidea): A nocturnal 290(5806): 470–474. doi:10.1038/290470a0. PMID:7219536. butterfly with superposition optics. Arthropod. Struct. Dev. Parsons, M. 1999. Butterflies of New Guinea. Academic Press, 36: 11–22. doi:10.1016/j.asd.2006.07.001. PMID:18089084. London, United Kingdom. Zhang, J., Cong, Q., Fan, X.L., Wang, R., Wang, M., and Preston-Mafham, K. 1991. Madagascar, a natural history. Facts Grishin, N.V. 2017a. Mitogenomes of Giant-Skipper Butter- on file Inc., Oxford, United Kingdom and New York. flies reveal an ancient split between deep and shallow root

Genome Downloaded from www.nrcresearchpress.com by Nick V. Grishin on 03/06/20 Sahoo, R.K., Warren, A.D., Wahlberg, N., Brower, A.V., feeders. F1000Res. 6: 222. doi:10.12688/f1000research.10970.1. Lukhtanov, V.A., and Kodandaramaiah, U. 2016. Ten genes PMID:28408977. and two topologies: an exploration of higher relationships in Zhang, J., Cong, Q., Shen, J., Fan, X.-L., Wang, M., and skipper butterflies (Hesperiidae). PeerJ, 4: e2653. doi:10.7717/ Grishin, N.V. 2017b. The complete mitogenome of Euschemon peerj.2653. PMID:27957386. rafflesia (Lepidoptera: Hesperiidae). Mitochondrial DNA Scoble, M.J. 1986. The structure and affinities of the Hedyloidea: Part B, 2(1): 136–138. doi:10.1080/23802359.2017.1292478. a new concept of the butterflies. Bull Br. Mus. 53(5): 251–286. PMID:30931394. Shao, L., Sun, Q., and Hao, J. 2015. The complete mitochondrial Zhang, J., Cong, Q., Shen, J., Brockmann, E., and Grishin, N.V. genome of Parara guttata (Lepidoptera: Hesperiidae). Mito- 2019. Three new subfamilies of skipper butterflies (Lepidop- chondrial DNA, 26(5): 724–725. doi:10.3109/19401736.2013. tera, Hesperiidae). Zookeys, 861: 91–105. doi:10.3897/zookeys. 845759. PMID:24460163. 861.34686. PMID:31333327.

Published by NRC Research Press